Passive Polarimetric Reconstruction Of Extended Dipole Target

We present a novel method for passive radar that simultaneously reconstructs the scene reflectivity and polarimetric states of stationary targets. Our method uses a spatially sparse distribution of polarimetrically diverse receivers to measure the backscattered signal of a scene illuminated by a source of opportunity. Our data model explicitly accounts for polarization and anisotropy of the target which is inherent in the multistatic configuration. We assume that each receiver is equipped with a pair of orthogonally polarized antennas, and form data as the pairwise correlation of the signal measured at different receivers. This results in the data being a linear mapping of the tensor product between two three-dimensional vector valued functions which represent the reflectivity and polarmetric states of the target. This tensor product can be represented as an unknown rank-1 operator with matrix-valued kernel. After discretization, this unknown operator can be modeled as an 4 th order tensor. We approach recovery of this unknown tensor from an optimization framework, exploiting its known structure. We demonstrate the performance of our approach with numerical simulations, and observe improved performance over the generalized likelihood ratio test approach.